AI agents retrieve information, interpret it, create new content, interact with systems, and act on a user’s behalf. That changes the data security conversation.

For years, security leaders have worked to understand where sensitive data lives, who can access it, and how to prevent inappropriate use or disclosure. Those challenges have not disappeared. In fact, agentic AI adds speed, scale, and new dependencies to an already difficult task.

AI Agents Magnify Existing Data Weaknesses That Sit Outside Of Security’s Core Responsibilities

Agentic AI exposes and amplifies the data weaknesses organizations already have. Unclear classification, uncertain ownership, inconsistent permissions, incomplete lineage, and questionable data quality become more consequential when an agent can reach across repositories and has increasing autonomy to act.

Conversations about inspecting prompts or blocking sensitive responses are necessary but insufficient. What makes securing data a challenge with agentic AI is that security does not own every decision that determines whether data is accurate, appropriate, properly retained, or suitable for an AI use case. Data governance and management practices, privacy stakeholders, data leadership, and business requirements all play a part in establishing AI-ready data. AI adoption will not pause while organizations attempt to resolve years of data management and governance debt to achieve this goal.

With AI In The Mix, How Do We Proceed To Protect Our Data?

Forrester’s AEGIS Framework for securing agentic AI outlines eight controls within its Data Security And Privacy domain as necessary for protecting data. These controls are not all technical controls. They include a mix of process and practices that require alignment with data, privacy, and AI engineering stakeholders to achieve across agent and data control design.

Security teams typically focus first on AEGIS-DATA-02 (data classification) within this domain. Organizations see the value of more consistent classification and better inventories before moving forward. The key is to move forward from here and not be haunted by the ghosts of past DLP deployments. Today, data security (tightly coupled with identity context for access and authorization!) requires the ability to enforce policies for AI agent access and interactions with data, and subsequent downstream actions.

Security leaders need a practical way to distinguish between the risks they can address directly where dependencies make autonomous behavior too risky, and the risks that require broader alignment with data and privacy functions where each of these functions will have varying levels of maturity too.

Join me at Forrester’s Security & Risk Forum for my closing keynote “Are We There Yet, Data Security?” where I’ll examine what security leaders must do to move forward today to protect data while taking steps to address dependencies in parallel. See you in Washington D.C. Nov 9 and 10!

Share